(Supplementary Information)

Table of Contents

Figure S1: [Chromatograms for SEC and SAX separations for three enzymatic digested CS] ... 2

Figure S2: [13C-gHSQC NMR analyses of chromatography enriched CS hexasaccharides ΔC4;4;4S-ol, ΔC6;6;6S-ol, C6;4;0S-ol and ΔC4;2,6;6S-ol] ...... 3

Figure S3: [MS spectra for reduced saturated trisulfated hexasaccharides with and without derivatizations] ...... 4

Figure S4: [RPLC chromatograms for three mixtures of CS trisulfated hexasaccharides from hyaluronidase digested CS-A] ...... 5

Structural analysis of CS disulfated and tetrasulfated hexasaccharides ...... 6

Figure S5: [Sulfation site identification of disulfated and tetra-sulfated hexasaccharides] ...... 9

Figure S6: [MS2 of [M+Na]+ (m/z 556.2) for derivatized unreduced disaccharide standards] ....10

Figure S1. Chromatograms for SEC (a–c) and SAX separations (d-f). Three enzymatic digested CS samples subjected to SEC fractionation are showed including C lyase digested CS-A (a), C lyase digested CS-C (b) and hyaluronidase digested CS-A (c). The numbers on the top of each peak correspond to the number of monosaccharide residues. The SAX chromatograms illustrate hexasaccharides with specific sulfation patterns purified from hexasaccharide fractions from each SEC separation: ΔC6;6;6S-ol and ΔC4;4;4S-ol from C lyase digested CS-A (d), ΔC4;2,6;6S-ol from C lyase digested CS-C (e), and C6;4;0S-ol from hyaluronidase digested CS-A (f).

Figure S2.13C-gHSQC NMR analyses of chromatography enriched CS hexasaccharides ΔC4;4;4S-ol (a), ΔC6;6;6S-ol (b), C6;4;0S-ol (c),andΔC4;2,6;6S-ol (d).The cross-peaks, which denote the sulfation sites, are indicated respectively in the NMR spectra. The nre-, mid-, and -ol stand for GalNAc units positioned at the nonreducing end, middle, and reducing terminal, respectively. The nre-4S peak in (d)has lower intensity due to pre-saturation effects at peaks nearby the water signal.

Figure S3. MS spectra for reduced saturated trisulfated hexasaccharides with and without derivatizations. Spectra are all full mass scans obtained under FT mode in negative ion mode for underivatized (a) and permethylated (b) products, and in positive ion mode for permethylated-desulfated (c) and permethylated-desulfated-acetylated (d) products. (a) [M – 3H]3–at m/z = 464.74 gives molecular weight of 1397.22. (b) [M – 3H]3–at m/z = 548.84 gives molecular weight of 1649.52 indicating 18 permethylation sites including the hydroxyl groups of the carboxyls on the GlcA and the hydrogen of the acetamine group on the GalNAc, leaving the sulfate groups unmodified. Less than 5% of the products are under-permethylated with 14 Da mass off (m/z of 544.16). Other significant peaks are disaccharide (558.19), tetrasaccharide (535.15, 543.17 and 551.16), or hexasaccharide (538.16) byproducts caused by the β-elimination reaction during permethylation, which are not selected for MSn analysis. (c) [M+2Na]2+ at m/z = 727.82 gives molecular weight of 1409.64 indicating all three sulfate groups are desulfated with less than 10% partial desulfated products (m/z of 767.82). (d) [M+2Na]2+ at m/z = 790.84 gives molecular weight of 1535.68 and no ion is detected around m/z of 769.83, indicating all three desulfated sites are fully acetylated. The final under-permethylated product that may be misidentified as a more highly sulfated species is detected at m/z of 804.84 at very low abundance.

Figure S4. RPLC chromatograms for three mixtures of CS trisulfated hexasaccharides from hyaluronidase-digested CS-A. The single ion chromatogram (SIC) for each mixture withdifferent derivatizations is extracted with corresponding parent ions: (a) SIC of [M+2Na]2+ (m/z 727.8) for the permethylated, desulfated mixture. (b) SIC of [M+2Na]2+ (m/z 757.8) for the permethylated, desulfated, trideuteromethylated mixture. (c) SIC of [M+2Na]2+ (m/z 790.8) for the permethylated, desulfated, acetylated mixture. Previous SAX purification and NMR analyses indicate three major trisulfated hexasaccharide products in this mixture.

Structural Analysis of CS Disulfated and Tetrasulfated Hexasaccharides

To validate the current methodology for recognition of sulfation patterns in CS oligosaccharides with differential sulfation degrees, two other CS saturated hexasaccharides were derivatized and analyzed: the disulfated C6;4;0S-ol and the tetra-sulfated ΔC4;2,6;6S-ol, both also characterized previously by NMR. While the successful identifications of the monosulfated disaccharide units, A-unit and C-unit, with either unsaturated or saturated nonreducing end were performed as described above, nonsulfated unit (O-unit) and disulfated units (D-unit and E-unit) were not tested. These units give a characteristic mass shift compared withthe monosulfated A- and C-units for their relative parent and product ions as listed in Table 1, and thus can easily differentiate monosulfated disaccharide units based on the disaccharide unit product ion masses. The mass shift caused by changes of sulfation degrees are calculated and listed in Table 1, where shifts of 28 Da are caused by the mass difference between the acetyl group and the methyl group. Hexasaccharides with one more sulfation group would generate derivatized products with one more acetyl group along with one less methyl group. This causes the increase of 28 Da for the product ions containing the sulfation site. Similarly, hexasaccharides with one less sulfation group would have a 28 Da less m/z value for relative ions. Between the derivatized oligosaccharides with saturated or unsaturated nonreducing end, a mass difference of 32 Da (MeOH) was observed for the parent ion and the [B2+Na]+ ion, but not for the other product ions that do not contain the nonreducing end. In addition, the [C1+Na]+ ion was only observed for saturated species (Figure 4a), while the [B20,2X5+Na]+ ion was only observed for unsaturated species (Figure 3a).

Following the fragmentation strategy shown in Figure 2a, we could easily target the product ions having mass shifts and therefore identify the degrees of sulfation for each disaccharide. The MS/MS spectra of hexasaccharides containing O-unit (C6;4;0S-ol) and D-unit (ΔC4;2,6;6S-ol) are shown in Figure S5. MS3 of [Y4+Na]+ for C6;2;0S-ol (Figure S5a) produced the [Y2+Na]+ ion with m/z of 534.2 having 28 Da less than the [Y2+Na]+ ion (m/z 562.2) shown in Figure 2c, indicating that the disaccharide at the reducing end bears no sulfation. For the tetra-sulfated ΔC4;2,6;6S-ol, the MS4 of [B4Y4+Na]+ (Figure S5b) gave the product ion [B4Z3+Na]+ with m/z of 278.2, indicating that the internal GalNAc is sulfated at the 6-position instead of the 4-position, while the [C3Y4+Na]+ ion (m/z 287.2) is 28 Da more massive than the one shown in Figure 3e(m/z 259.2), indicating the internal GlcA was sulfated. Thus, the MSn data indicated the internal disaccharide unit of ΔC4;2,6;6S-ol is disulfated, where one sulfation is on the GalNAc unit and the other sulfation is on the GlcA unit, distinguishing the D-unit from the E-unit (GlcA-GalNAc4,6S) in CS oligosaccharides.

We have shown that the methodology presented here can be successfully used for identification of CS hexasaccharides containing four different types of disaccharide units except for the hexasaccharides with E-unit, which is currently absent in our CS hexameric library. While the differentiation of D-unit and E-unit should be easily obtained by observing the mass difference on the appropriate GlcA and/or GalNAc, we still derivatized and analyzed two disulfated disaccharide standards, the ΔC2,6S and the ΔC46S, to demonstrate the ability of the derivatization scheme to efficiently distinguish the E-unit from the D-unit, and to show the differences of the MS/MS spectra between these two CS disaccharides. The ΔC2,6S was prepared in the same way as hexasaccharides preparation, while the ΔC46S was purchased from Dextra (Reading, UK) and analyzed by NMR to verify the structure of the commercial standard (data not shown). Both disaccharides were permethylated, desulfated, and acetylated without previous reduction in order to mimic an MSn product ion not from the reducing end of an oligosaccharide. The MS2 spectra of [M+Na]+ (m/z 556.2) of both CS dimers are shown in Figure S6. The m/z for Z1 ion observed for ΔC2,6S was 310.2 (16 Da less than the Z1 ion shown in Table 1due to the unreduced reducing end) indicating only one sulfation site on the GalNAc residue. The Z1 ion for ΔC46S, on the other hand, with a m/z of 338.2 was 28 Da more than the Z1 ion for ΔC2,6S,indicating two sulfated sites on GalNAc. The 0,2X1 ion, a cross-ring cleavage fragment including carbons C1 and C2 of the GlcA and GalNAc, gave the same m/z value (412.2) for both disaccharides showing that the second sulfated site for ΔC2,6S was on the C2 position of GlcA. Thus, the results indicate that our methodology can easily differentiate the D-unit from the E-unit, supporting the assignment of CS hexamers bearing any sulfation type at any position.

Figure S5. Sulfation site identification of disulfated and tetrasulfated hexasaccharides:(a) MS3 of [Y4+Na]+ (m/z 1025.4) for derivatized C6;4;0S-ol. The m/z for Y2 ion observed here is 534.2, which is 28 mass units less than 562.2, indicating the GalNAc on the reducing end is nonsulfated. (b) MS4 of [B4Y4+Na]+ (m/z 542.2) for derivatized ΔC4;2,6;6S-ol. The m/z for C3Y4 ion observed here is 287.2, which is 28 mass units more than 259.2, indicating the internal GlcA is sulfated.

Figure S6. MS2 of [M+Na]+ (m/z 556.2) for derivatized unreduced disaccharide standards: ΔC2,6S (a) and ΔC46S (b). The m/z for Z1 ion observed for ΔC2,6S is 310.2 which is 28 Da less than 338.2 for ΔC46S, indicating one less sulfated site on GalNAc. The 0,2X1 ion, a cross-ring cleavage fragment ion including carbons C1 and C2 of the GlcA, gives the same m/z value (412.2) showing that the second sulfated site for ΔC2,6S is on the C2 position of GlcA.

Page 1 of 10